On public capital analysis with state data

Based on state production models with fixed effects, recent studies have argued public capital is not productive. We show multicollinearity is a potential problem and caution is warranted in interpreting estimation results for models with public capital and fixed effects.